INDEX

Explanations

phrases indicating a list of items or ideas

oai_token-act-pair · gpt-3.5-turbo

the recurring phrase "of" in various 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

Not in Any Lists

No Comments

Negative Logits

dayName

-0.73

 agre

-0.70

WER

-0.67

allery

-0.66

iture

-0.63

ertodd

-0.63

anse

-0.62

cms

-0.60

MpServer

-0.60

IPM

-0.59

POSITIVE LOGITS

arching

0.73

sson

0.69

icial

0.69

 course

0.67

 whom

0.65

sted

0.61

ten

0.60

 Thousand

0.58

 Arrest

0.58

Ľ

0.58

Activations Density 0.083%

phrases indicating a list of items or ideas

the recurring phrase "of" in various contexts

No Comments

No Known Activations