INDEX

Explanations

phrases related to sequential events or actions

oai_token-act-pair · gpt-3.5-turbo

frequent use of the word "Again" and its associated context in discussions or narratives

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

emo

-0.64

 Defender

-0.64

 nevertheless

-0.63

esm

-0.62

 nonetheless

-0.61

 Cooldown

-0.61

aila

-0.59

iasis

-0.58

 Built

-0.57

bon

-0.57

POSITIVE LOGITS

dor

0.75

ItemImage

0.73

ADRA

0.71

azeera

0.67

 dilig

0.67

 earthqu

0.66

IFIED

0.66

ortmund

0.65

ierre

0.65

orie

0.65

Activations Density 0.088%

phrases related to sequential events or actions

frequent use of the word "Again" and its associated context in discussions or narratives

No Comments

No Known Activations