INDEX

Explanations

terms related to specific technical details or concepts

oai_token-act-pair · gpt-3.5-turbo

references to various crises and related issues

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

ĸļ

-0.83

bsite

-0.68

emaker

-0.58

anwhile

-0.57

usher

-0.56

terday

-0.56

ynthesis

-0.56

leneck

-0.55

forcer

-0.55

cale

-0.55

POSITIVE LOGITS

 variants

1.00

 types

0.94

 varieties

0.94

 locations

0.94

 albums

0.94

 cases

0.93

coins

0.93

 positions

0.92

 beers

0.91

 titles

0.90

Activations Density 0.783%

terms related to specific technical details or concepts

references to various crises and related issues

No Comments

No Known Activations