INDEX

Explanations

references to counting or quantifying

oai_token-act-pair · gpt-3.5-turbo

the term "Count" and its variations related to numerical grouping or classification

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

 wrench

-0.68

 embr

-0.66

 appre

-0.65

Â·Â·

-0.62

 shel

-0.62

 conduc

-0.61

 alarms

-0.60

 tears

-0.58

BIT

-0.58

 tram

-0.58

POSITIVE LOGITS

ries

1.31

ies

1.29

ess

1.21

eenth

1.18

een

1.13

esses

1.10

enance

1.09

reys

1.04

less

1.01

down

1.00

Activations Density 0.044%

references to counting or quantifying

the term "Count" and its variations related to numerical grouping or classification

No Comments

No Known Activations