INDEX

Explanations

political or legal terms related to regulations and rulings

oai_token-act-pair · gpt-3.5-turbo

components related to interference and claims

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

 nightmares

-0.68

 tremend

-0.61

 haunted

-0.61

 knights

-0.60

houses

-0.59

nexus

-0.59

 envy

-0.58

ashtra

-0.57

Ô

-0.55

annis

-0.55

POSITIVE LOGITS

achment

0.91

asion

0.86

ciating

0.84

erences

0.84

etary

0.81

incial

0.81

andise

0.81

essional

0.80

ruction

0.80

ession

0.79

Activations Density 0.099%

political or legal terms related to regulations and rulings

components related to interference and claims

No Comments

No Known Activations