INDEX

Explanations

patterns of characters that appear at the end of a string

oai_token-act-pair · gpt-3.5-turbo

abbreviations or acronyms related to measurements or media terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Saras

-0.91

 Hancock

-0.85

Hus

-0.80

 Sonny

-0.77

bush

-0.75

shr

-0.75

 Goth

-0.74

 Savannah

-0.74

 Indra

-0.73

-0.72

POSITIVE LOGITS

et

1.88

ET

1.70

ett

1.56

ets

1.50

Ã©t

1.47

ETS

1.41

ete

1.35

etry

1.35

etting

1.35

etts

1.33

Activations Density 0.333%

patterns of characters that appear at the end of a string

abbreviations or acronyms related to measurements or media terminology

No Comments

No Known Activations