INDEX

Explanations

words related to dishonest behavior such as lying and cheating

oai_token-act-pair · gpt-3.5-turbo

words related to dishonesty or deception

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

00200000

-0.70

DragonMagazine

-0.68

oran

-0.68

erick

-0.68

ä½ľ

-0.66

osaurs

-0.64

ership

-0.63

shown

-0.63

oval

-0.63

nesota

-0.61

POSITIVE LOGITS

 cheating

0.98

 cheat

0.94

 sheet

0.79

cheat

0.78

heet

0.75

 cheated

0.71

 sheets

0.70

 spouses

0.68

 testers

0.68

bowl

0.68

Activations Density 0.040%

words related to dishonest behavior such as lying and cheating

words related to dishonesty or deception

No Comments

No Known Activations