INDEX

Explanations

references to technology product series

oai_token-act-pair · gpt-3.5-turbo

mentions of product series or model numbers

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

bled

-0.84

bered

-0.75

arching

-0.74

selves

-0.74

adel

-0.72

bats

-0.71

rette

-0.69

prus

-0.69

bling

-0.69

inently

-0.67

POSITIVE LOGITS

CI

0.81

manship

0.75

GP

0.75

ctive

0.75

ĸļ

0.74

BuyableInstoreAndOnline

0.74

Series

0.71

CRIP

0.70

Graph

0.69

andi

0.69

Activations Density 0.028%

references to technology product series

mentions of product series or model numbers

No Comments

No Known Activations