INDEX

Explanations

the word "weird" and its variations

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "weirdness."

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

ptive

-0.85

ptives

-0.84

vation

-0.79

adr

-0.77

apers

-0.77

ILA

-0.76

aders

-0.75

HI

-0.74

REL

-0.73

ailable

-0.72

POSITIVE LOGITS

ness

1.05

nesses

0.92

ly

0.91

entimes

0.85

 Weird

0.83

 weird

0.83

oes

0.79

est

0.79

ishly

0.78

 occurrences

0.75

Activations Density 0.014%

the word "weird" and its variations

references to the concept of "weirdness."

No Comments

No Known Activations