INDEX

Explanations

proper nouns or locations containing the word "Point"

oai_token-act-pair · gpt-3.5-turbo

mentions of various locations referred to as "Point" followed by a number

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

ccording

-0.87

 delinqu

-0.81

ESE

-0.78

 eleph

-0.77

 chemically

-0.76

 destro

-0.76

 metabol

-0.75

 notor

-0.72

thy

-0.71

versely

-0.70

POSITIVE LOGITS

 Point

1.25

point

1.09

Point

1.06

eous

0.86

points

0.84

 Forward

0.84

Points

0.81

 Bree

0.80

hole

0.77

pad

0.77

Activations Density 0.010%

proper nouns or locations containing the word "Point"

mentions of various locations referred to as "Point" followed by a number

No Comments

No Known Activations