INDEX

Explanations

proper nouns followed by specific patterns of characters

oai_token-act-pair · gpt-3.5-turbo

references to news sources and categories of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

pse

-0.56

 vanishing

-0.54

 score

-0.51

onite

-0.50

 Â¶

-0.50

 sshd

-0.50

vom

-0.50

 clock

-0.49

manship

-0.48

 chew

-0.48

POSITIVE LOGITS

meier

0.56

igion

0.55

Friend

0.54

Reference

0.53

Value

0.53

Hon

0.53

 Exploration

0.52

Commerce

0.51

ada

0.51

Wr

0.51

Activations Density 0.142%

proper nouns followed by specific patterns of characters

references to news sources and categories of information

No Comments

No Known Activations