INDEX

Explanations

terms relating to uncertainty or lack of knowledge

oai_token-act-pair · gpt-3.5-turbo

the repeated use of the word "how" 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

UME

-0.73

odder

-0.68

agonists

-0.67

izu

-0.63

ature

-0.62

ware

-0.61

 Issue

-0.61

iculture

-0.60

 ........

-0.59

lehem

-0.59

POSITIVE LOGITS

soever

0.81

HCR

0.81

 much

0.78

beit

0.77

ls

0.76

ever

0.75

 MUCH

0.74

©¶æ

0.73

ling

0.72

much

0.71

Activations Density 0.081%

terms relating to uncertainty or lack of knowledge

the repeated use of the word "how" in various contexts

No Comments

No Known Activations