INDEX

Explanations

technical terms related to science or specific locations mentioned without proper context

oai_token-act-pair · gpt-3.5-turbo

symbols and specialized notations

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

wagen

-0.93

uster

-0.92

nut

-0.89

iqueness

-0.88

nuts

-0.86

gow

-0.86

gren

-0.86

wcs

-0.84

stro

-0.82

matic

-0.82

POSITIVE LOGITS

âĶĢâĶĢ

0.94

âĨĴ

0.90

 âĨĴ

0.87

ãĥĺ

0.87

âĶĢâĶĢâĶĢâĶĢ

0.86

Â·Â·

0.82

 âĪ

0.81

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.80

âĢ¢âĢ¢âĢ¢âĢ¢

0.78

âĨĳ

0.77

Activations Density 0.005%

technical terms related to science or specific locations mentioned without proper context

symbols and specialized notations

No Comments

No Known Activations