INDEX

Explanations

mentions of new surveys or reports

oai_token-act-pair · gpt-3.5-turbo

references to new surveys and reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

cius

-0.93

 actionGroup

-0.88

Zip

-0.82

usercontent

-0.75

omever

-0.75

ufact

-0.75

 Himself

-0.73

agree

-0.71

ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ

-0.71

uddin

-0.70

POSITIVE LOGITS

bie

1.15

bies

0.96

 testament

0.94

 batch

0.94

 YORK

0.91

 arrivals

0.90

foundland

0.89

 additions

0.88

 edition

0.88

 revelations

0.87

Activations Density 0.087%

mentions of new surveys or reports

references to new surveys and reports

No Comments

No Known Activations