INDEX

Explanations

instances where something is demonstrated or presented

oai_token-act-pair · gpt-3.5-turbo

instances of the word "showing" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

tymology

-0.83

undai

-0.82

etheless

-0.81

rit

-0.78

cult

-0.78

perty

-0.72

orrow

-0.71

rafted

-0.68

lake

-0.67

tar

-0.67

POSITIVE LOGITS

Hide

0.91

shows

0.87

imony

0.83

-+-+

0.81

 attRot

0.81

 SHOW

0.80

Redditor

0.79

Show

0.78

 Shows

0.77

ered

0.77

Activations Density 0.009%

instances where something is demonstrated or presented

instances of the word "showing" in various contexts

No Comments

No Known Activations