INDEX

Explanations

names of people in a list format

oai_token-act-pair · gpt-3.5-turbo

names of people and characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 glim

-0.66

 NETWORK

-0.62

Reviewer

-0.62

 FIRE

-0.58

 bulletin

-0.55

 contradictory

-0.52

 tremend

-0.52

 toget

-0.51

 pleas

-0.51

 FANTASY

-0.51

POSITIVE LOGITS

)|

0.66

Jr

0.62

 respectively

0.60

ona

0.60

alli

0.59

]]

0.59

uty

0.58

fur

0.57

aux

0.56

osa

0.56

Activations Density 1.550%

names of people in a list format

names of people and characters

No Comments

No Known Activations