INDEX

Explanations

titular words in titles or phrases

oai_token-act-pair · gpt-3.5-turbo

titles or references to named entities and specific designations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ITNESS

-0.69

shapeshifter

-0.68

 saline

-0.65

net

-0.65

 Observer

-0.64

 procedural

-0.64

 gravity

-0.62

 magnetic

-0.61

 downhill

-0.60

POSITIVE LOGITS

Tit

1.11

itles

1.06

ename

0.85

resy

0.84

reon

0.84

iago

0.81

indal

0.80

arus

0.80

zek

0.80

orius

0.77

Activations Density 0.004%

titular words in titles or phrases

titles or references to named entities and specific designations

No Comments

No Known Activations