INDEX

Explanations

cryptic symbols or codes that appear within the text

oai_token-act-pair · gpt-3.5-turbo

numerical data and sequences

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

 tremend

-1.02

 eleph

-0.90

 citiz

-0.85

aditional

-0.83

 occas

-0.80

senal

-0.79

ò

-0.79

oun

-0.78

 exting

-0.77

 undermin

-0.77

POSITIVE LOGITS

↵

0.97

Merit

0.67

ÂłÂłÂł

0.67

cms

0.65

Round

0.64

assembly

0.63

Jere

0.63

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.63

Chel

0.62

Frames

0.62

Activations Density 0.370%

cryptic symbols or codes that appear within the text

numerical data and sequences

No Comments

No Known Activations