INDEX

Explanations

Roman numerals

oai_token-act-pair · gpt-3.5-turbo

references to the Roman numeral "II" and its various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

=-=-=-=-=-=-=-=-

-0.80

ãĥ¼ãĥĨ

-0.79

ãĤ±

-0.73

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.71

ãĤ©

-0.70

acters

-0.70

ãĥ£

-0.68

Ð´

-0.68

ãĤ§

-0.67

rooms

-0.66

POSITIVE LOGITS

II

1.20

III

1.17

HF

0.91

FY

0.89

ATA

0.82

III

0.78

pec

0.76

ND

0.76

HS

0.75

HK

0.73

Activations Density 0.012%

Roman numerals

references to the Roman numeral "II" and its various contexts

No Comments

No Known Activations