INDEX

Explanations

words related to darkness or negative contexts

oai_token-act-pair · gpt-3.5-turbo

references to darkness or negative connotations associated with the term "dark."

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

cule

-0.67

kson

-0.67

yip

-0.67

agine

-0.66

JUST

-0.62

 Canaver

-0.61

Pok

-0.60

available

-0.60

Dur

-0.59

 Vital

-0.57

POSITIVE LOGITS

ening

1.28

room

1.06

ened

1.06

horse

0.93

rooms

0.92

skinned

0.91

vision

0.89

net

0.88

 skies

0.87

haired

0.85

Activations Density 0.025%

words related to darkness or negative contexts

references to darkness or negative connotations associated with the term "dark."

No Comments

No Known Activations