INDEX

Explanations

the word "Good" preceded by another word

oai_token-act-pair · gpt-3.5-turbo

the word "Good" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

âĹ¼

-1.09

RAW

-0.77

-0.73

eters

-0.73

oths

-0.70

eds

-0.68

EStream

-0.66

Tumblr

-0.65

Lago

-0.65

ople

-0.65

POSITIVE LOGITS

bye

1.16

enough

1.11

 Enough

0.97

bye

0.93

 Good

0.92

good

0.92

reads

0.87

Bye

0.87

luck

0.86

 Luck

0.85

Activations Density 0.011%

the word "Good" preceded by another word

the word "Good" in various contexts

No Comments

No Known Activations