INDEX

Explanations

phrases related to intensification or extremeness

oai_token-act-pair · gpt-3.5-turbo

mentions of being "full" or "new" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ensation

-0.65

-0.62

cephal

-0.61

peat

-0.61

ovember

-0.60

-0.58

cause

-0.58

clock

-0.58

terday

-0.58

ocide

-0.57

POSITIVE LOGITS

 Alloy

0.57

 Cody

0.56

 Joel

0.54

 Camel

0.54

 Rocky

0.53

 Colorado

0.52

 Pharma

0.52

 Reyes

0.51

Vir

0.51

 Norris

0.51

Activations Density 1.406%

phrases related to intensification or extremeness

mentions of being "full" or "new" in various contexts

No Comments

No Known Activations