INDEX

Explanations

instances where a particular "thing" is emphasized within a context

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing important or notable statements

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

inav

-0.76

ULT

-0.72

cul

-0.71

oÄŁ

-0.70

lems

-0.70

DOS

-0.67

bush

-0.67

ONSORED

-0.67

JV

-0.65

vez

-0.65

POSITIVE LOGITS

 happens

0.78

iverse

0.76

rued

0.74

 happened

0.70

 separates

0.68

 Valiant

0.65

afort

0.64

 counts

0.63

 happening

0.62

 touches

0.60

Activations Density 0.032%

instances where a particular "thing" is emphasized within a context

phrases emphasizing important or notable statements

No Comments

No Known Activations