INDEX

Explanations

negations or refutations of claims

oai_token-act-pair · gpt-3.5-turbo

negations related to various topics or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ãĤ¼ãĤ¦ãĤ¹

-0.72

eur

-0.69

çīĪ

-0.69

stakes

-0.69

itech

-0.66

¿½

-0.64

 Finder

-0.64

Org

-0.63

å¥

-0.63

 Films

-0.62

POSITIVE LOGITS

 adequately

1.17

 necessarily

1.17

icably

1.09

 sufficiently

0.97

 properly

0.94

epad

0.94

icable

0.94

hin

0.90

 assimil

0.89

 bother

0.89

Activations Density 0.228%

negations or refutations of claims

negations related to various topics or statements

No Comments

No Known Activations