INDEX

Explanations

phrases related to legal procedures, laws, and political discussions

oai_token-act-pair · gpt-3.5-turbo

references to legal and medical topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ãĢ

-0.80

âĸº

-0.74

 ðŁĳ

-0.70

âĢ

-0.69

@#&

-0.68

ãĢĲ

-0.67

 âĶĤ

-0.66

 Âł Âł Âł Âł Âł Âł Âł Âł

-0.65

 âĸº

-0.64

Æ

-0.64

POSITIVE LOGITS

 themselves

0.96

 merely

0.88

 outright

0.84

selves

0.83

 spontaneously

0.76

 respectively

0.76

 inadvertently

0.76

 altogether

0.75

 falsely

0.75

 selves

0.75

Activations Density 0.820%

phrases related to legal procedures, laws, and political discussions

references to legal and medical topics

No Comments

No Known Activations