INDEX

Explanations

sentence endings emphasizing negative aspects or consequences

oai_token-act-pair · gpt-3.5-turbo

sentences or phrases that express strong emotions or conclusions

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

Winged

-0.64

iden

-0.63

ãĥĭ

-0.63

®

-0.62

 Meier

-0.62

ictional

-0.62

 Leilan

-0.61

 restraining

-0.59

umat

-0.58

ighed

-0.58

POSITIVE LOGITS

Interstitial

0.96

Cause

0.88

tis

0.83

taboola

0.80

til

0.80

nel

0.77

emouth

0.76

cause

0.74

Mech

0.72

ny

0.70

Activations Density 0.026%

sentence endings emphasizing negative aspects or consequences

sentences or phrases that express strong emotions or conclusions

No Comments

No Known Activations