INDEX

Explanations

terms related to searching activities or actions

oai_token-act-pair · gpt-3.5-turbo

terms related to searching or looking for something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

mos

-0.73

ario

-0.68

Ģ

-0.67

apo

-0.62

eno

-0.60

pret

-0.60

Problem

-0.59

touch

-0.59

adra

-0.59

hire

-0.58

POSITIVE LOGITS

 frantically

1.03

for

1.01

 unsuccessfully

0.97

 diligently

0.94

 vain

0.91

 through

0.90

 tirelessly

0.86

 eBay

0.83

 Craigslist

0.81

 databases

0.79

Activations Density 0.115%

terms related to searching activities or actions

terms related to searching or looking for something

No Comments

No Known Activations