INDEX

Explanations

questions starting with "How can" or variations of this phrase

oai_token-act-pair · gpt-3.5-turbo

questions about possibilities and potential actions

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

 skating

-0.73

board

-0.71

 simulator

-0.65

 cruising

-0.65

boarding

-0.64

 threaded

-0.62

Board

-0.62

 subscribed

-0.62

RAFT

-0.60

cart

-0.59

POSITIVE LOGITS

't

1.12

berra

1.09

ADA

0.88

iencies

0.79

aries

0.78

adian

0.78

sts

0.77

ary

0.75

ataka

0.73

itic

0.71

Activations Density 0.051%

questions starting with "How can" or variations of this phrase

questions about possibilities and potential actions

No Comments

No Known Activations