INDEX

Explanations

capitalized words that may indicate titles or headings

oai_token-act-pair · gpt-3.5-turbo

references to 'reviews' or 'rating' systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

nib

-0.67

ANE

-0.66

ONES

-0.65

 Yuan

-0.64

 Zamb

-0.64

gag

-0.63

 Slim

-0.63

 Nest

-0.62

 simulator

-0.62

NEC

-0.61

POSITIVE LOGITS

porter

1.35

asons

1.22

ceiver

1.18

becca

1.16

cover

1.15

leases

1.15

venge

1.14

gex

1.13

joice

1.12

putable

1.11

Activations Density 0.021%

capitalized words that may indicate titles or headings

references to 'reviews' or 'rating' systems

No Comments

No Known Activations