INDEX

Explanations

phrases related to different levels of achievement or measurement

oai_token-act-pair · gpt-3.5-turbo

references to varying levels of proficiency or performance

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

vous

-0.81

rontal

-0.74

ãĥĹ

-0.66

 Terra

-0.65

Arg

-0.65

 Phot

-0.64

 Andromeda

-0.63

lez

-0.63

 retina

-0.62

Wil

-0.62

POSITIVE LOGITS

level

1.03

 level

0.99

 levels

0.93

Level

0.90

levels

0.80

 Level

0.77

afety

0.72

 Levels

0.72

rise

0.72

ropri

0.69

Activations Density 0.030%

phrases related to different levels of achievement or measurement

references to varying levels of proficiency or performance

No Comments

No Known Activations