INDEX

Explanations

the prefix "non-" in words

oai_token-act-pair · gpt-3.5-turbo

references to the term "non" 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Tycoon

-0.95

 Orchestra

-0.79

 Oaks

-0.77

 Memories

-0.77

 Spoon

-0.77

 Sparrow

-0.76

 Showdown

-0.76

 Grind

-0.76

 McGee

-0.76

 Grill

-0.75

POSITIVE LOGITS

etheless

1.02

zero

1.01

chal

1.00

fiction

0.99

epad

0.99

aligned

0.96

basic

0.93

availability

0.93

resident

0.92

linear

0.91

Activations Density 0.021%

the prefix "non-" in words

references to the term "non" in various contexts

No Comments

No Known Activations