INDEX

Explanations

words related to the prefix "ce-", likely focusing on terms related to change or leadership

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "cessation" or related terms

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

rition

-1.05

bucks

-0.96

ested

-0.90

wagen

-0.90

*/(

-0.84

rats

-0.80

 Cosponsors

-0.80

stadt

-0.79

STER

-0.79

rider

-0.76

POSITIVE LOGITS

Ce

1.06

ce

0.97

ce

0.96

asers

0.91

plin

0.82

cca

0.81

cc

0.79

resp

0.77

asar

0.76

 Dame

0.75

Activations Density 0.007%

words related to the prefix "ce-", likely focusing on terms related to change or leadership

references to the concept of "cessation" or related terms

No Comments

No Known Activations