INDEX

Explanations

the intention or capability to perform an action

oai_token-act-pair · gpt-3.5-turbo

phrases indicating potential actions or capabilities of subjects

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

 irony

-0.64

 Wrestling

-0.63

 treacher

-0.62

 understatement

-0.62

ARK

-0.59

 significance

-0.58

wheel

-0.58

 hypocrisy

-0.57

 agitation

-0.57

addons

-0.57

POSITIVE LOGITS

sylv

1.00

 could

0.81

can

0.80

mie

0.77

ctic

0.75

'll

0.74

 maximize

0.74

 qualify

0.72

arers

0.70

 comply

0.70

Activations Density 0.151%

the intention or capability to perform an action

phrases indicating potential actions or capabilities of subjects

No Comments

No Known Activations