INDEX

Explanations

mentions of fines or penalties in legal contexts

oai_token-act-pair · gpt-3.5-turbo

references to financial penalties or fines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

omorphic

-0.96

omorph

-0.93

ational

-0.75

ween

-0.74

ocaly

-0.73

ãĤ¤ãĥĪ

-0.72

OB

-0.68

pres

-0.68

ococ

-0.66

itals

-0.66

POSITIVE LOGITS

 levied

0.96

 fine

0.88

fine

0.78

 fines

0.76

 Gael

0.73

Fine

0.71

 tuning

0.69

 evasion

0.67

($

0.66

ishment

0.66

Activations Density 0.011%

mentions of fines or penalties in legal contexts

references to financial penalties or fines

No Comments

No Known Activations