INDEX

Explanations

strings of characters that resemble machine code or other technical formatting symbols

oai_token-act-pair · gpt-3.5-turbo

symbols or special characters associated with legal or judicial contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

•GPT2 L5 Resid: RRLLRLLRLLRRRRLLRRRRRRRRR

No Comments

Negative Logits

 scattering

-0.77

bye

-0.75

 coll

-0.71

 scatter

-0.71

 sails

-0.70

 dispers

-0.70

 Voyager

-0.70

 platinum

-0.68

 glasses

-0.68

 barrier

-0.67

POSITIVE LOGITS

Į

1.00

MORE

0.94

ulty

0.93

Attempts

0.93

¹

0.93

£

0.92

¤

0.89

CBC

0.87

Update

0.86

ł

0.85

Activations Density 0.180%

strings of characters that resemble machine code or other technical formatting symbols

symbols or special characters associated with legal or judicial contexts

No Comments

No Known Activations