INDEX

Explanations

geographical locations, specifically named regions

oai_token-act-pair · gpt-3.5-turbo

references to the name "Far" in various contexts

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

sburgh

-0.91

ettings

-0.79

vironment

-0.76

ichick

-0.75

ysis

-0.67

İĭ

-0.67

essee

-0.67

 Kinn

-0.63

 Socrates

-0.63

ificial

-0.62

POSITIVE LOGITS

riers

0.92

rier

0.91

ouk

0.87

az

0.87

rak

0.86

aday

0.85

thing

0.85

rer

0.84

ghan

0.83

rug

0.83

Activations Density 0.015%

geographical locations, specifically named regions

references to the name "Far" in various contexts

No Comments

No Known Activations