INDEX

Explanations

names containing the pattern "Kr" with varying levels of specificity

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Krzanich" and its variations

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

 Dragonbound

-0.81

ãģ¦

-0.70

 Samoa

-0.68

 Flickr

-0.65

quo

-0.65

BOOK

-0.63

eering

-0.63

 Canadiens

-0.62

å§«

-0.61

ORGE

-0.60

POSITIVE LOGITS

usk

1.16

assic

1.15

istani

1.03

uder

0.99

asus

0.99

©¶æ¥µ

0.96

ota

0.94

acy

0.94

aska

0.93

unk

0.92

Activations Density 0.011%

names containing the pattern "Kr" with varying levels of specificity

mentions of the name "Krzanich" and its variations

No Comments

No Known Activations