INDEX

Explanations

dates and years in a specific format

oai_token-act-pair · gpt-3.5-turbo

dates, specifically years in the late 1980s

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

uten

-0.99

udic

-0.89

isters

-0.83

eter

-0.79

eton

-0.78

maid

-0.77

fix

-0.75

indal

-0.73

jriwal

-0.73

antry

-0.72

POSITIVE LOGITS

 Cumm

0.80

omial

0.71

Ñĭ

0.71

ï¿½ï¿½ï¿½ï¿½

0.70

0.69

 Ferdinand

0.69

rities

0.68

 Scotia

0.67

âĸĪ

0.66

é¾

0.65

Activations Density 0.040%

dates and years in a specific format

dates, specifically years in the late 1980s

No Comments

No Known Activations