INDEX

Explanations

mentions of and details related to cigarettes

oai_token-act-pair · gpt-3.5-turbo

references to cigarettes and smoking

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

herty

-0.81

*/(

-0.79

pmwiki

-0.77

isite

-0.75

variable

-0.73

Reviewer

-0.69

UFC

-0.68

icter

-0.68

ithmetic

-0.67

 Yates

-0.67

POSITIVE LOGITS

arette

1.31

arettes

1.24

 cigarettes

1.13

 smoking

1.09

 smokers

1.05

 smoker

1.04

 cigarette

1.02

cig

1.00

cigarettes

1.00

 smoked

1.00

Activations Density 0.012%

mentions of and details related to cigarettes

references to cigarettes and smoking

No Comments

No Known Activations