INDEX

Explanations

terms related to events or actions that happened before a specific point in time

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "prior."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

aden

-0.79

asp

-0.70

tower

-0.69

owship

-0.69

hack

-0.67

 elves

-0.65

=-=-=-=-=-=-=-=-

-0.65

ickle

-0.63

rosso

-0.63

tre

-0.61

POSITIVE LOGITS

itiz

1.11

itized

1.09

etheless

0.87

ities

0.85

ebin

0.82

cies

0.80

icip

0.77

emort

0.75

 generations

0.73

 authenticated

0.71

Activations Density 0.011%

terms related to events or actions that happened before a specific point in time

mentions of the term "prior."

No Comments

No Known Activations