INDEX

Explanations

words related to a specific concept or title, "Way"

oai_token-act-pair · gpt-3.5-turbo

references to "Way," often relating to philosophical or directional concepts

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

 ceilings

-0.65

ickets

-0.65

 suppress

-0.65

ãĥĥãĤ¯

-0.64

quart

-0.63

 sensit

-0.63

hips

-0.63

 reciproc

-0.59

 recip

-0.59

pless

-0.59

POSITIVE LOGITS

Way

3.87

Way

2.32

WAY

2.11

way

1.73

 Ways

1.52

way

1.49

WAY

1.35

 Highway

1.18

 Path

1.17

 Kind

1.06

Activations Density 0.008%

words related to a specific concept or title, "Way"

references to "Way," often relating to philosophical or directional concepts

No Comments

No Known Activations