INDEX

Explanations

words related to a specific person or concept, likely named "Buff"

oai_token-act-pair · gpt-3.5-turbo

references to a character named "Buff" or variations of that name

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

________________________

-0.80

FTWARE

-0.68

 Labyrinth

-0.63

PF

-0.62

PLAY

-0.62

————————

-0.61

 Stras

-0.60

QUIRE

-0.59

 Worlds

-0.59

ACTION

-0.58

POSITIVE LOGITS

alon

1.13

alos

1.07

ered

1.06

enstein

0.97

ett

0.94

dinand

0.94

ering

0.90

 Buff

0.90

erey

0.90

alo

0.89

Activations Density 0.013%

words related to a specific person or concept, likely named "Buff"

references to a character named "Buff" or variations of that name

No Comments

No Known Activations