INDEX

Explanations

words related to specific companies, organizations, or products

oai_token-act-pair · gpt-3.5-turbo

nouns related to cultural or artistic concepts

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

 ..."

-0.64

.</

-0.64

]."

-0.58

 Harvey

-0.58

 âĢ¦"

-0.56

 [âĢ¦]

-0.55

âĢ¦.

-0.55

.;

-0.54

âĢ¦."

-0.54

âĢ¦..

-0.54

POSITIVE LOGITS

ulhu

0.82

hester

0.78

raltar

0.72

ragon

0.70

osate

0.67

idia

0.67

culosis

0.66

obook

0.66

eport

0.66

miah

0.65

Activations Density 0.372%

words related to specific companies, organizations, or products

nouns related to cultural or artistic concepts

No Comments

No Known Activations