INDEX

Explanations

specific named entities related to a place or person

oai_token-act-pair · gpt-3.5-turbo

references to the word "Par" or variations of it in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ħĭ

-0.81

sonian

-0.73

éĹĺ

-0.68

 Carbuncle

-0.68

 dwar

-0.67

å£«

-0.67

OME

-0.66

=]

-0.66

ĻĤ

-0.64

ARDS

-0.64

POSITIVE LOGITS

agraph

1.29

anoia

1.24

vati

1.19

allel

1.15

alysis

1.11

cel

1.06

agon

1.06

amount

1.04

abolic

1.02

cells

1.01

Activations Density 0.019%

specific named entities related to a place or person

references to the word "Par" or variations of it in different contexts

No Comments

No Known Activations