INDEX

Explanations

a variety of different scenarios or possibilities described in the text

oai_token-act-pair · gpt-3.5-turbo

phrases related to hypothetical or worst-case scenarios

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

enfranch

-0.80

ances

-0.76

cipline

-0.75

inters

-0.72

ateurs

-0.72

vals

-0.70

aina

-0.70

roots

-0.70

algia

-0.69

obe

-0.68

POSITIVE LOGITS

 scenario

1.60

 scenarios

1.53

enario

1.04

 whereby

0.96

 involving

0.95

 unfolding

0.95

 unfold

0.94

 envis

0.94

 wherein

0.94

 envisioned

0.93

Activations Density 0.091%

a variety of different scenarios or possibilities described in the text

phrases related to hypothetical or worst-case scenarios

No Comments

No Known Activations