INDEX

Explanations

scientific terms or alphanumeric combinations like symbols

oai_token-act-pair · gpt-3.5-turbo

the occurrences of a specific symbol or character and its variations in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

uster

-0.96

wagen

-0.92

enic

-0.88

iqueness

-0.88

gow

-0.86

igree

-0.86

nut

-0.85

eni

-0.84

isf

-0.84

nuts

-0.84

POSITIVE LOGITS

 âĨĴ

1.06

âĶĢâĶĢ

0.99

âĨĴ

0.97

âĶĢâĶĢâĶĢâĶĢ

0.95

 âĨ

0.91

âĢ¢âĢ¢âĢ¢âĢ¢

0.89

 âĪ

0.85

Â·Â·

0.83

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.82

 ãĤµ

0.82

Activations Density 0.008%

scientific terms or alphanumeric combinations like symbols

the occurrences of a specific symbol or character and its variations in text

No Comments

No Known Activations