INDEX

Explanations

phrases indicating an action, decision, or result

oai_token-act-pair · gpt-3.5-turbo

instances of strong affirmative or negative assertions in relation to events or conditions

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

 nonetheless

-0.76

»Ĵ

-0.72

cheat

-0.69

etheless

-0.67

reply

-0.67

ssh

-0.66

 disg

-0.65

loo

-0.63

rect

-0.61

 Madness

-0.61

POSITIVE LOGITS

ãĥ¯

0.68

 Rowe

0.64

INA

0.61

}{

0.59

ãģ®éŃĶ

0.58

DERR

0.58

urally

0.58

 guiActive

0.58

 aesthetics

0.57

 guiActiveUn

0.57

Activations Density 0.199%

phrases indicating an action, decision, or result

instances of strong affirmative or negative assertions in relation to events or conditions

No Comments

No Known Activations