INDEX

Explanations

mentions or references to people with the last name "Gilligan" along with a high activation value for some mentions of the word "offense"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "igan" in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ngth

-0.76

grades

-0.74

graded

-0.73

dates

-0.72

grading

-0.69

iles

-0.68

nesday

-0.67

ĺħ

-0.65

 spaced

-0.64

ensions

-0.64

POSITIVE LOGITS

igans

1.53

igan

1.41

 Nights

0.87

omics

0.87

 Athletic

0.85

cheon

0.81

ews

0.80

osa

0.80

owski

0.79

wolf

0.78

Activations Density 0.018%

mentions or references to people with the last name "Gilligan" along with a high activation value for some mentions of the word "offense"

occurrences of the term "igan" in various contexts

No Comments

No Known Activations