INDEX

Explanations

phrases indicating certainty or emphasis

oai_token-act-pair · gpt-3.5-turbo

instances of negation or refusal in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

RAD

-0.51

 guiActiveUnfocused

-0.51

creen

-0.49

 scattering

-0.48

 shroud

-0.48

 cottage

-0.46

 lodging

-0.46

 Manhattan

-0.46

 scatter

-0.46

 shack

-0.46

POSITIVE LOGITS

¬

0.82

£

0.78

¡

0.77

¹

0.77

Ĵ

0.77

ı

0.74

¼

0.74

º

0.73

Â¢

0.71

ł

0.69

Activations Density 0.516%

phrases indicating certainty or emphasis

instances of negation or refusal in various contexts

No Comments

No Known Activations