INDEX

Explanations

information or discussions related to private conversations or actions

oai_token-act-pair · gpt-3.5-turbo

mentions of private matters or confidentiality

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

 GOODMAN

-0.95

 Surge

-0.83

=-=-=-=-=-=-=-=-

-0.80

rers

-0.76

amins

-0.74

EVA

-0.74

 Ashe

-0.73

rer

-0.72

estones

-0.72

rament

-0.70

POSITIVE LOGITS

 conclud

0.88

 owned

0.85

 drilled

0.82

 repro

0.79

 confess

0.78

 privately

0.78

UGE

0.77

 disposed

0.77

 insured

0.76

 rebuilt

0.75

Activations Density 0.012%

information or discussions related to private conversations or actions

mentions of private matters or confidentiality

No Comments

No Known Activations