INDEX

Explanations

the word "second" or similar variations

oai_token-act-pair · gpt-3.5-turbo

the word "second" and its variations, suggesting a focus on sequential or additional points being made

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

Sab

-0.59

 Rapt

-0.57

Bright

-0.56

entious

-0.56

anned

-0.55

 mats

-0.55

-0.53

own

-0.52

ga

-0.52

Jou

-0.52

POSITIVE LOGITS

 secondly

0.93

 Gutenberg

0.80

ngth

0.80

named

0.77

itto

0.75

 importantly

0.73

ividual

0.71

isSpecialOrderable

0.69

ername

0.69

etheless

0.67

Activations Density 0.221%

the word "second" or similar variations

the word "second" and its variations, suggesting a focus on sequential or additional points being made

No Comments

No Known Activations