INDEX

Explanations

negative statements or negations

oai_token-act-pair · gpt-3.5-turbo

negations and the phrase "isn't" 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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

tein

-0.74

gnu

-0.66

 creations

-0.60

 Properties

-0.58

 endeavors

-0.58

 Supported

-0.57

doms

-0.55

 WARN

-0.54

 pursuits

-0.54

ritch

-0.53

POSITIVE LOGITS

hin

1.01

ibaba

1.00

 anybody

1.00

 enough

0.99

 unanim

0.98

 anyone

0.98

 anything

0.97

any

0.96

enough

0.93

 room

0.90

Activations Density 0.079%

negative statements or negations

negations and the phrase "isn't" in various contexts

No Comments

No Known Activations