INDEX

Explanations

phrases related to physical structures such as doors and windows

oai_token-act-pair · gpt-3.5-turbo

references to doors and their symbolism or functional roles 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 @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

utan

-0.75

owicz

-0.71

omic

-0.66

Summary

-0.66

 summarized

-0.64

FF

-0.62

 Supplement

-0.62

atum

-0.61

Analysis

-0.59

arse

-0.59

POSITIVE LOGITS

 doors

4.12

 door

2.63

 Doors

2.50

doors

2.46

 gates

2.22

door

2.10

 Door

2.07

 curtains

1.79

 entrances

1.75

 windows

1.65

Activations Density 0.017%

phrases related to physical structures such as doors and windows

references to doors and their symbolism or functional roles in various contexts

No Comments

No Known Activations