INDEX

Explanations

instances of the word "ols" with varying activation levels

oai_token-act-pair · gpt-3.5-turbo

mentions of "pistols" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ŃĶ

-0.72

 Rapp

-0.70

office

-0.66

liest

-0.62

FACE

-0.58

 Pentagon

-0.58

 French

-0.57

Chair

-0.57

 Kenyan

-0.57

PRESS

-0.56

POSITIVE LOGITS

ols

1.39

olics

1.04

ength

1.00

terday

0.99

ongs

0.91

olic

0.90

atile

0.90

ands

0.89

ipop

0.88

allery

0.85

Activations Density 0.010%

instances of the word "ols" with varying activation levels

mentions of "pistols" in various contexts

No Comments

No Known Activations