INDEX

Explanations

the word "don't" in various contexts

oai_token-act-pair · gpt-3.5-turbo

negations or negative phrasing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 enthusi

-0.81

 newcom

-0.72

anwhile

-0.71

 reluct

-0.70

senal

-0.68

£ı

-0.66

 hemor

-0.65

 rall

-0.65

 exha

-0.65

 challeng

-0.64

POSITIVE LOGITS

't

1.61

ned

1.04

gered

0.90

ning

0.87

iencies

0.86

itely

0.86

uts

0.83

ÃŃ

0.82

oit

0.81

UT

0.81

Activations Density 0.093%

the word "don't" in various contexts

negations or negative phrasing

No Comments

No Known Activations