INDEX

Explanations

references to physical force or power

oai_token-act-pair · gpt-3.5-turbo

references to "force" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

algia

-0.81

alam

-0.75

vironment

-0.73

ecause

-0.71

STER

-0.71

roma

-0.70

Hop

-0.69

artment

-0.69

andon

-0.69

DERR

-0.68

POSITIVE LOGITS

maj

1.29

 Awakens

0.99

 exerted

0.99

 multiplier

0.93

fulness

0.87

 force

0.84

vic

0.84

ps

0.79

 induction

0.75

 multipl

0.75

Activations Density 0.044%

references to physical force or power

references to "force" in various contexts

No Comments

No Known Activations