INDEX

Explanations

names of specific places or institutions

oai_token-act-pair · gpt-3.5-turbo

mentions of places or organizations, particularly those related to 'Ow' or 'Ll'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

bourg

-0.85

 centrif

-0.71

 gravity

-0.62

gravity

-0.61

 unres

-0.61

ãĥ¼ãĥĨ

-0.61

itol

-0.60

 fracturing

-0.59

 inaccur

-0.58

 congr

-0.58

POSITIVE LOGITS

ow

1.42

Owl

1.11

owl

1.11

swick

1.06

yssey

1.03

ahoo

1.01

chemy

1.00

ls

0.99

orld

0.95

tale

0.91

Activations Density 0.010%

names of specific places or institutions

mentions of places or organizations, particularly those related to 'Ow' or 'Ll'

No Comments

No Known Activations