INDEX

Explanations

instances where the word "aha" is mentioned

oai_token-act-pair · gpt-3.5-turbo

the repeated expression of laughter or amusement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lines

-0.75

lining

-0.71

master

-0.70

paid

-0.70

cutting

-0.66

part

-0.65

writing

-0.64

rations

-0.63

 dime

-0.62

writers

-0.61

POSITIVE LOGITS

aha

1.38

ahah

1.04

ibaba

0.94

kson

0.89

uthor

0.88

hea

0.87

gee

0.85

HAHA

0.85

izabeth

0.83

oche

0.82

Activations Density 0.004%

instances where the word "aha" is mentioned

the repeated expression of laughter or amusement

No Comments

No Known Activations