INDEX

Explanations

numbers (including years) combined with words related to historical events

oai_token-act-pair · gpt-3.5-turbo

numeric values and identifiers associated with statistics or data

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Flavoring

-0.64

tering

-0.62

amac

-0.61

iman

-0.58

dict

-0.56

lins

-0.56

stration

-0.56

 Situation

-0.55

 brill

-0.54

mission

-0.54

POSITIVE LOGITS

1.71

1.67

1.66

1.61

1.57

1.52

1.51

1.43

1.16

Activations Density 0.123%

numbers (including years) combined with words related to historical events

numeric values and identifiers associated with statistics or data

No Comments

No Known Activations