INDEX

Explanations

expressions related to different aspects or aspects of something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "fronts."

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

Negative Logits

utan

-0.90

app

-0.61

utor

-0.59

gyn

-0.59

imate

-0.59

 award

-0.59

 accident

-0.58

eka

-0.58

asm

-0.57

Pic

-0.57

POSITIVE LOGITS

 fronts

4.32

 facets

1.20

front

1.16

 lobb

1.16

 angles

1.15

 avenues

1.15

 bases

1.09

 pillars

1.08

 corners

1.06

 foundations

1.06

Activations Density 0.008%

expressions related to different aspects or aspects of something

instances of the word "fronts."

No Known Activations