INDEX

Explanations

words related to beliefs, presumptions, hypotheses, or propositions

oai_token-act-pair · gpt-3.5-turbo

phrases related to beliefs or hypotheses 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

Interstitial

-0.86

sung

-0.83

waters

-0.74

odes

-0.69

incinn

-0.67

HCR

-0.65

oho

-0.64

umen

-0.64

adan

-0.64

crew

-0.64

POSITIVE LOGITS

 assumption

1.04

 assumptions

1.01

 underpin

0.80

 assumes

0.78

 Lauder

0.75

staking

0.74

 incorrectly

0.70

 disclaimer

0.70

 assum

0.68

 premise

0.68

Activations Density 0.011%

words related to beliefs, presumptions, hypotheses, or propositions

phrases related to beliefs or hypotheses in various contexts

No Comments

No Known Activations