INDEX

Explanations

verbs or phrases related to negative or problematic situations

oai_token-act-pair · gpt-3.5-turbo

negative outcomes or failures

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

 Dome

-0.65

ilion

-0.62

aurus

-0.58

 pursu

-0.58

 Tart

-0.58

 Demand

-0.58

 Sapp

-0.55

zo

-0.54

Advertisements

-0.53

Hok

-0.53

POSITIVE LOGITS

Ĥİ

0.92

ŃĶ

0.80

¿½

0.78

osate

0.76

inka

0.70

ĸļ

0.67

Ĥª

0.67

glers

0.66

¨

0.66

 partName

0.66

Activations Density 0.610%

verbs or phrases related to negative or problematic situations

negative outcomes or failures

No Comments

No Known Activations