INDEX

Explanations

names of specific investigations or reports

oai_token-act-pair · gpt-3.5-turbo

colons that introduce content or lists

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

xton

-0.69

pell

-0.63

lag

-0.63

 tremend

-0.61

 renown

-0.61

tsky

-0.59

 branch

-0.59

gee

-0.57

 decoration

-0.56

 terminating

-0.55

POSITIVE LOGITS

 Cosponsors

0.81

 Provided

0.76

NPR

0.72

%]

0.68

 Scores

0.66

atican

0.66

taboola

0.64

ispers

0.64

Fact

0.62

Study

0.61

Activations Density 0.039%

names of specific investigations or reports

colons that introduce content or lists

No Comments

No Known Activations