INDEX

Explanations

words related to something very obvious, clear, and often negative

oai_token-act-pair · gpt-3.5-turbo

instances of the word "blatant" and its variations, indicating obvious or clear wrongdoing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

agine

-0.87

ovember

-0.82

arios

-0.78

yip

-0.77

ulton

-0.77

nesota

-0.74

iership

-0.74

erves

-0.70

ivas

-0.69

 safely

-0.68

POSITIVE LOGITS

 disregard

0.99

 misrepresent

0.94

 hypocrisy

0.93

iary

0.91

 violations

0.88

 violation

0.86

 blatant

0.83

 fals

0.79

 disreg

0.79

 falsehood

0.78

Activations Density 0.040%

words related to something very obvious, clear, and often negative

instances of the word "blatant" and its variations, indicating obvious or clear wrongdoing

No Comments

No Known Activations