INDEX

Explanations

phrases or terms related to individuals or personalities

oai_token-act-pair · gpt-3.5-turbo

references to various individuals or characters

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

DERR

-0.77

CCC

-0.66

 Crusade

-0.65

osterone

-0.64

TPP

-0.62

 Ministers

-0.62

mud

-0.60

 Protestants

-0.59

 Belt

-0.58

 surge

-0.58

POSITIVE LOGITS

hood

1.46

nel

1.33

ification

1.21

nels

1.18

age

1.12

ified

1.12

ality

1.09

ages

1.08

ifies

1.06

ifications

1.04

Activations Density 0.073%

phrases or terms related to individuals or personalities

references to various individuals or characters

No Comments

No Known Activations