INDEX

Explanations

patterns of numbers and letters in a specific sequence

oai_token-act-pair · gpt-3.5-turbo

sequences of numeric values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

é¾

-0.93

geries

-0.90

taboola

-0.86

externalActionCode

-0.83

CVE

-0.81

èª

-0.77

glomer

-0.75

 fortun

-0.75

conservancy

-0.74

clinton

-0.73

POSITIVE LOGITS

0.80

Qi

0.80

 Guards

0.76

 Unsure

0.76

Ara

0.76

0.75

 âľĶ

0.75

 Onion

0.74

 Yoshi

0.73

 Wool

0.72

Activations Density 0.065%

patterns of numbers and letters in a specific sequence

sequences of numeric values

No Comments

No Known Activations