INDEX

Explanations

dates formatted as months, days, and years

oai_token-act-pair · gpt-3.5-turbo

closing parentheses in text

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

ãĥ¥

-0.75

 artif

-0.71

buck

-0.69

sing

-0.67

 shorth

-0.67

 bait

-0.65

 audible

-0.65

 answ

-0.64

 watch

-0.64

ause

-0.63

POSITIVE LOGITS

 srfAttach

0.84

>]

0.81

 Committees

0.76

ONSORED

0.72

//[

0.71

âķ

0.70

ATURE

0.69

Origin

0.68

eem

0.68

 Panc

0.68

Activations Density 0.117%

dates formatted as months, days, and years

closing parentheses in text

No Comments

No Known Activations