INDEX

Explanations

references to jokes

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "jokes."

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

ignty

-0.69

CLASSIFIED

-0.67

ioch

-0.63

 violet

-0.62

 Borders

-0.61

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.60

orneys

-0.60

itizen

-0.58

JUST

-0.58

ignt

-0.57

POSITIVE LOGITS

 jokes

1.02

linger

0.81

ters

0.80

 joking

0.78

ster

0.78

 banter

0.78

sters

0.76

ong

0.75

osal

0.74

itone

0.74

Activations Density 0.010%

references to jokes

occurrences of the word "jokes."

No Comments

No Known Activations