INDEX

Explanations

phrases or terms known by other names

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate alternative names or descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 fixation

-0.71

udic

-0.70

nexus

-0.63

avorite

-0.62

osion

-0.62

adobe

-0.61

ftime

-0.60

vette

-0.59

users

-0.59

efficiency

-0.58

POSITIVE LOGITS

 gener

0.76

 acronym

0.75

 simply

0.73

literally

0.73

as

0.73

 pseudonym

0.73

 derog

0.72

 collectively

0.72

 abbrevi

0.72

 abbre

0.71

Activations Density 0.054%

phrases or terms known by other names

phrases that indicate alternative names or descriptions

No Comments

No Known Activations