INDEX

Explanations

the word "won't"

oai_token-act-pair · gpt-3.5-turbo

negations or expressions of refusal

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

Kings

-0.66

Must

-0.66

 Communities

-0.65

Liter

-0.65

ancer

-0.65

Intern

-0.65

 Strategy

-0.65

Aren

-0.65

 Measures

-0.64

Pure

-0.63

POSITIVE LOGITS

 necessarily

1.22

bud

1.02

 bother

1.02

icably

1.00

be

0.92

 spoil

0.89

 hesitate

0.89

 tolerate

0.87

 exactly

0.87

icable

0.81

Activations Density 0.050%

the word "won't"

negations or expressions of refusal

No Comments

No Known Activations