INDEX

Explanations

phrases related to arguments or explanations used to justify or support various viewpoints

oai_token-act-pair · gpt-3.5-turbo

phrases that involve the concept of defense or defending something

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

*/(

-0.69

--------------------------------------------------------

-0.68

phrine

-0.65

SOURCE

-0.64

together

-0.63

------------------------------------------------

-0.63

ucket

-0.62

hook

-0.60

 baskets

-0.59

 Around

-0.59

POSITIVE LOGITS

 sanity

0.81

tnc

0.76

 dignity

0.76

 integrity

0.75

 indef

0.73

ilege

0.70

itely

0.69

 Freed

0.68

 endangered

0.68

 preservation

0.68

Activations Density 0.182%

phrases related to arguments or explanations used to justify or support various viewpoints

phrases that involve the concept of defense or defending something

No Comments

No Known Activations