INDEX

Explanations

mathematically related terms or concepts, such as calculations or physical arrangements with specific attributes

oai_token-act-pair · gpt-3.5-turbo

action words related to composition and arrangement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

aucus

-0.75

 Depot

-0.74

arty

-0.73

idated

-0.72

arthy

-0.71

umbn

-0.71

ixels

-0.66

å§«

-0.66

port

-0.66

cdn

-0.66

POSITIVE LOGITS

river

0.89

BY

0.82

by

0.72

 solely

0.70

abl

0.69

 nowhere

0.67

ocument

0.64

 exclusively

0.64

aw

0.64

nesday

0.64

Activations Density 0.271%

mathematically related terms or concepts, such as calculations or physical arrangements with specific attributes

action words related to composition and arrangement

No Comments

No Known Activations