INDEX

Explanations

phrases related to legal or investigative inquiries

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate relationships and connections to various topics or issues

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

LI

-0.82

©¶æ¥µ

-0.81

SUP

-0.74

vier

-0.73

icer

-0.71

herer

-0.70

iem

-0.70

 encount

-0.68

Written

-0.67

¨

-0.67

POSITIVE LOGITS

 mater

0.71

 behalf

0.67

 whether

0.66

the

0.66

 matters

0.65

 differing

0.65

 geography

0.64

 Species

0.63

 subsistence

0.63

 topics

0.62

Activations Density 0.095%

phrases related to legal or investigative inquiries

phrases that indicate relationships and connections to various topics or issues

No Comments

No Known Activations