INDEX

Explanations

specific names or terms containing the word "lan"

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names related to individuals or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 correction

-0.73

 decom

-0.66

 doct

-0.66

 acceptance

-0.65

 release

-0.64

 upgrade

-0.64

 throw

-0.63

 unchanged

-0.62

 collection

-0.61

 restoration

-0.60

POSITIVE LOGITS

lan

4.77

LAN

1.84

lane

1.84

len

1.57

lin

1.38

lon

1.37

lan

1.34

la

1.30

lins

1.22

dan

1.14

Activations Density 0.006%

specific names or terms containing the word "lan"

proper nouns, specifically names related to individuals or entities

No Comments

No Known Activations