INDEX

Explanations

the phrase "you know," specifically highlighting information or personal insights

oai_token-act-pair · gpt-3.5-turbo

instances of repetition or affirmation in conversation

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

erity

-0.85

oslav

-0.82

¬¼

-0.78

omal

-0.76

Init

-0.71

mage

-0.70

oreal

-0.70

onding

-0.69

nai

-0.69

ĸļ

-0.69

POSITIVE LOGITS

uh

0.90

sir

0.68

 WHERE

0.67

 CLASS

0.64

 Harriet

0.64

hed

0.63

LET

0.63

 tick

0.62

 gentlemen

0.61

 WHAT

0.60

Activations Density 0.041%

the phrase "you know," specifically highlighting information or personal insights

instances of repetition or affirmation in conversation

No Comments

No Known Activations