INDEX

Explanations

family-related words, especially "mom."

oai_token-act-pair · gpt-3.5-turbo

nouns and terms related to organizational or structural elements

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

åĮ

-0.74

Compat

-0.70

NetMessage

-0.66

 undermin

-0.65

Semitism

-0.64

$.

-0.63

abwe

-0.62

 alike

-0.62

 Yorkers

-0.61

Lago

-0.60

POSITIVE LOGITS

iest

1.02

liest

0.95

 analogy

0.81

 hypothesis

0.79

 portion

0.75

est

0.72

 aspect

0.69

 scanner

0.69

 aisle

0.67

 angle

0.65

Activations Density 0.847%

family-related words, especially "mom."

nouns and terms related to organizational or structural elements

No Comments

No Known Activations