INDEX

Explanations

organizations or entities related to defense, security, intelligence, and technology

oai_token-act-pair · gpt-3.5-turbo

references to governmental or defense-related entities and terminology

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

å§«

-0.91

natureconservancy

-0.86

é¾įåĸļå£«

-0.83

imensional

-0.78

historic

-0.75

urat

-0.71

mallow

-0.70

urnal

-0.69

å¥³

-0.68

advertising

-0.67

POSITIVE LOGITS

AMA

0.63

otiation

0.59

letters

0.59

 symb

0.58

 Giul

0.56

backs

0.56

olutely

0.55

linger

0.55

 entitlement

0.55

 startup

0.55

Activations Density 0.578%

organizations or entities related to defense, security, intelligence, and technology

references to governmental or defense-related entities and terminology

No Comments

No Known Activations