INDEX

Explanations

phrases indicating a decision or action believed to be morally or practically correct

oai_token-act-pair · gpt-3.5-turbo

discussions about moral or ethical decisions

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

culosis

-0.72

irie

-0.67

inder

-0.64

issy

-0.64

 unsus

-0.62

qus

-0.60

gian

-0.59

ravings

-0.59

 dominance

-0.58

 lodging

-0.58

POSITIVE LOGITS

 choices

0.81

 financially

0.80

 forward

0.78

for

0.73

 politically

0.69

nered

0.67

 choice

0.66

 mate

0.64

 chosen

0.64

 ideally

0.64

Activations Density 0.169%

phrases indicating a decision or action believed to be morally or practically correct

discussions about moral or ethical decisions

No Comments

No Known Activations