INDEX

Explanations

references to the term "FA" with different numerical values

oai_token-act-pair · gpt-3.5-turbo

references to "FA" as it relates to a sporting context

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

atchewan

-0.69

 Reson

-0.69

owsky

-0.67

 Lizard

-0.64

tie

-0.63

 Drift

-0.62

 Horton

-0.61

 Amen

-0.60

selves

-0.60

come

-0.59

POSITIVE LOGITS

ctions

0.95

ruits

0.95

ULTS

0.94

FA

0.90

ITH

0.90

ILS

0.89

ZE

0.88

uthor

0.87

ederation

0.87

BIP

0.84

Activations Density 0.011%

references to the term "FA" with different numerical values

references to "FA" as it relates to a sporting context

No Comments

No Known Activations