INDEX

Explanations

words related to overlaying or superimposing one thing on another

oai_token-act-pair · gpt-3.5-turbo

terms related to overlapping concepts or layers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Reviewer

-0.73

 Sabha

-0.72

 Roma

-0.71

ãģ®éŃĶ

-0.71

ELY

-0.69

Charge

-0.67

RON

-0.65

culosis

-0.63

ãĥ¼ãĥĨ

-0.62

 Crimean

-0.61

POSITIVE LOGITS

 overl

1.03

ords

0.99

 overlay

0.95

ays

0.95

apped

0.94

ancies

0.88

odied

0.85

aying

0.84

atcher

0.78

aced

0.77

Activations Density 0.015%

words related to overlaying or superimposing one thing on another

terms related to overlapping concepts or layers

No Comments

No Known Activations