INDEX

Explanations

discussions related to holding law enforcement accountable

oai_token-act-pair · gpt-3.5-turbo

references to admissions or revelations in a narrative context

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

ordable

-0.79

 ÂŃ

-0.78

 affordability

-0.73

WASHINGTON

-0.72

rastructure

-0.72

terday

-0.72

Canada

-0.71

 Marketplace

-0.71

Federal

-0.70

markets

-0.69

POSITIVE LOGITS

 characters

0.95

 resemb

0.94

 Characters

0.92

 trope

0.91

 canon

0.90

Trivia

0.87

canon

0.87

 implied

0.85

 backstory

0.84

 narration

0.84

Activations Density 1.559%

discussions related to holding law enforcement accountable

references to admissions or revelations in a narrative context

No Comments

No Known Activations