INDEX

Explanations

references to the term "Fore" or similar keywords in various contexts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Fore" and its variations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

naires

-0.87

å°Ĩ

-0.83

Magikarp

-0.79

iating

-0.73

ä»£

-0.73

tumblr

-0.72

arsity

-0.70

Redditor

-0.70

uality

-0.68

Els

-0.67

POSITIVE LOGITS

 Fore

1.22

closure

1.01

sts

1.01

closed

0.93

runner

0.92

 fore

0.88

nsics

0.86

stal

0.85

pred

0.83

fore

0.83

Activations Density 0.009%

references to the term "Fore" or similar keywords in various contexts

occurrences of the word "Fore" and its variations

No Comments

No Known Activations