INDEX

Explanations

mentions of the term "obs" or related variations

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "obs."

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

Negative Logits

 Monarch

-0.75

 slip

-0.66

 Poles

-0.66

ãĥĥãĥī

-0.66

 Serbia

-0.64

WAY

-0.63

 Sakuya

-0.63

 Italy

-0.63

ca

-0.62

 historic

-0.61

POSITIVE LOGITS

obs

1.35

ols

1.06

essed

1.03

poon

1.01

ushi

0.99

ugar

0.97

ilib

0.97

ullivan

0.96

essions

0.96

afety

0.94

Activations Density 0.005%

mentions of the term "obs" or related variations

occurrences of the word "obs."

No Known Activations