INDEX

Explanations

the word "brick" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to bricks and brick-related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ntil

-0.77

ktop

-0.77

judicial

-0.71

selection

-0.69

 Voters

-0.69

opausal

-0.67

inen

-0.64

constitutional

-0.63

PsyNetMessage

-0.62

 relativity

-0.62

POSITIVE LOGITS

bats

1.27

layer

1.06

 bricks

0.97

mort

0.97

shaw

0.95

works

0.95

 brick

0.93

yard

0.91

buster

0.88

 Brick

0.87

Activations Density 0.007%

the word "brick" in various contexts

references to bricks and brick-related concepts

No Comments

No Known Activations