INDEX

Explanations

occurrences of the word "a" followed by an adjective

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "doing something."

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

 Arri

-0.77

 Mastery

-0.70

 Returning

-0.68

 Sett

-0.66

 Cursed

-0.66

 Riding

-0.65

 fears

-0.65

 Defeat

-0.64

 Removal

-0.64

inventoryQuantity

-0.64

POSITIVE LOGITS

 homework

0.88

 grunt

0.84

 bidding

0.81

job

0.79

 justice

0.78

umerable

0.77

 thing

0.77

 chores

0.75

 injustice

0.74

 experiment

0.72

Activations Density 0.115%

occurrences of the word "a" followed by an adjective

references to the concept of "doing something."

No Comments

No Known Activations