INDEX

Explanations

references to manuals or formal documents

oai_token-act-pair · gpt-3.5-turbo

references to manuals and documentation related to diagnostic criteria

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 mash

-0.73

una

-0.67

 mathemat

-0.63

 association

-0.62

 conspiracy

-0.61

pared

-0.60

 acquaintance

-0.60

tymology

-0.58

ibel

-0.57

onna

-0.57

POSITIVE LOGITS

Interstitial

0.80

 Leban

0.75

odes

0.72

GoldMagikarp

0.71

 Roose

0.67

 Dalai

0.66

axies

0.65

henko

0.64

 ende

0.64

ensable

0.63

Activations Density 0.002%

references to manuals or formal documents

references to manuals and documentation related to diagnostic criteria

No Comments

No Known Activations