INDEX

Explanations

contractions

oai_token-act-pair · gpt-3.5-turbo

special characters and punctuation that may indicate direct speech or possessiveness

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

Not in Any Lists

No Comments

Negative Logits

 scoreboard

-0.69

 decimal

-0.69

 Slug

-0.68

 silhou

-0.67

 telesc

-0.66

 seiz

-0.65

 extravag

-0.65

 enchantment

-0.64

åº

-0.63

 decap

-0.62

POSITIVE LOGITS

very

0.91

shall

0.88

should

0.88

really

0.87

lege

0.86

yet

0.81

conduct

0.81

must

0.81

been

0.80

âĢ¢âĢ¢

0.79

Activations Density 0.398%

contractions

special characters and punctuation that may indicate direct speech or possessiveness

No Comments

No Known Activations