INDEX

Explanations

phrases related to expressing opinions or attributing thoughts to others

oai_token-act-pair · gpt-3.5-turbo

statements or claims made by individuals regarding various topics

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

riage

-0.66

ricks

-0.63

geries

-0.63

 Forward

-0.61

 Engineers

-0.61

 Quan

-0.60

arnaev

-0.59

endered

-0.59

 Quentin

-0.58

 Trigger

-0.58

POSITIVE LOGITS

ģĸ

0.76

©¶æ¥µ

0.72

 sorely

0.70

 dearly

0.68

itud

0.67

abouts

0.67

é¾į

0.67

·

0.66

 deems

0.66

0.65

Activations Density 0.369%

phrases related to expressing opinions or attributing thoughts to others

statements or claims made by individuals regarding various topics

No Comments

No Known Activations