INDEX

Explanations

words related to a specific geometric shape, particularly the word "pent"

oai_token-act-pair · gpt-3.5-turbo

references to "pentathlon" and related terms

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

 Hitman

-0.68

 Malf

-0.68

 Carbuncle

-0.66

 Milo

-0.66

 Mahar

-0.64

itability

-0.64

ãģĦ

-0.63

CTV

-0.63

ISSION

-0.63

 Referred

-0.63

POSITIVE LOGITS

 pent

1.06

iple

1.01

agon

1.00

etr

0.98

inct

0.94

raph

0.88

ium

0.85

agons

0.85

urous

0.84

iate

0.84

Activations Density 0.020%

words related to a specific geometric shape, particularly the word "pent"

references to "pentathlon" and related terms

No Comments

No Known Activations