INDEX

Explanations

different versions or iterations of something

oai_token-act-pair · gpt-3.5-turbo

references to different versions or adaptations of stories or narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

urers

-0.87

arers

-0.76

phasis

-0.75

ateurs

-0.74

ussion

-0.70

oji

-0.68

ering

-0.67

hya

-0.67

ennes

-0.67

erning

-0.66

POSITIVE LOGITS

 Fortune

0.66

 fortune

0.66

 events

0.61

Xan

0.61

 Christianity

0.60

 Thing

0.59

CAT

0.59

 Tomorrow

0.59

reality

0.59

 Flip

0.59

Activations Density 0.117%

different versions or iterations of something

references to different versions or adaptations of stories or narratives

No Comments

No Known Activations