INDEX

Explanations

phrases related to specific items or aspects of physical objects

oai_token-act-pair · gpt-3.5-turbo

connections to significant figures or concepts 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

acious

-0.72

oe

-0.62

vernment

-0.62

hene

-0.62

ãĥ©ãĥ³

-0.62

itatively

-0.61

 Facts

-0.59

otaur

-0.58

entious

-0.57

foundland

-0.56

POSITIVE LOGITS

 substituted

0.82

 intervene

0.78

 excluded

0.76

 predominant

0.74

 unaffected

0.73

 interven

0.72

 suffice

0.71

 abound

0.69

 intervened

0.69

 untouched

0.66

Activations Density 1.126%

phrases related to specific items or aspects of physical objects

connections to significant figures or concepts in various contexts

No Comments

No Known Activations