INDEX

Explanations

the word "don't" preceded by 'I'

oai_token-act-pair · gpt-3.5-turbo

statements expressing denial or rejection

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 newcom

-0.65

 misdem

-0.65

 princ

-0.63

 Mandatory

-0.61

 eleph

-0.61

anwhile

-0.61

 challeng

-0.60

 subur

-0.59

 populated

-0.59

 enthusi

-0.58

POSITIVE LOGITS

't

1.64

ned

1.11

Â´

1.05

ÃŃ

0.96

ovan

0.91

gered

0.85

ates

0.85

uts

0.85

nel

0.83

essee

0.83

Activations Density 0.082%

the word "don't" preceded by 'I'

statements expressing denial or rejection

No Comments

No Known Activations