INDEX

Explanations

adjectives ending in -ed or -ing

oai_token-act-pair · gpt-3.5-turbo

words related to the state of being restricted or confined

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Clicker

-0.58

vac

-0.58

 sequest

-0.56

 kicking

-0.54

 venue

-0.54

 RELEASE

-0.52

 Garfield

-0.52

 Funds

-0.51

EST

-0.51

 postp

-0.51

POSITIVE LOGITS

ned

4.76

ning

2.46

ners

1.77

ted

1.64

nings

1.53

nel

1.45

ny

1.36

nered

1.30

NING

1.29

med

1.28

Activations Density 0.010%

adjectives ending in -ed or -ing

words related to the state of being restricted or confined

No Comments

No Known Activations