INDEX

Explanations

proper nouns related to a specific location, specifically "Baton Rouge"

oai_token-act-pair · gpt-3.5-turbo

mentions of specific locations, particularly Baton Rouge

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PLIED

-0.78

WAR

-0.72

ski

-0.71

war

-0.71

committee

-0.70

sche

-0.70

present

-0.69

defined

-0.69

ULT

-0.68

enance

-0.67

POSITIVE LOGITS

 Baton

1.30

 Rouge

1.06

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

0.83

 teasp

0.81

hesda

0.78

emouth

0.78

©¶æ¥µ

0.77

ãĥ£

0.77

ignt

0.75

xes

0.73

Activations Density 0.007%

proper nouns related to a specific location, specifically "Baton Rouge"

mentions of specific locations, particularly Baton Rouge

No Comments

No Known Activations