INDEX

Explanations

references to specific names, probably related to people or places

oai_token-act-pair · gpt-3.5-turbo

prominent nouns related to individuals or groups that play important roles

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

 âĢº

-0.48

 ..........

-0.48

 corpor

-0.47

-0.45

taboola

-0.44

----

-0.44

 Aram

-0.44

ĵĺ

-0.42

 Maver

-0.42

 Morty

-0.42

POSITIVE LOGITS

schild

0.52

ucer

0.45

erno

0.44

imore

0.43

usional

0.43

roller

0.42

ultimate

0.41

 Hera

0.41

enment

0.41

ahime

0.40

Activations Density 3.543%

references to specific names, probably related to people or places

prominent nouns related to individuals or groups that play important roles

No Comments

No Known Activations