INDEX

Explanations

instances where something is described as being a good or beneficial thing

oai_token-act-pair · gpt-3.5-turbo

indefinite articles and the general concept of "something" or "one" as it relates to 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agree

-0.85

Edit

-0.80

 Izan

-0.75

Versions

-0.72

ontent

-0.70

Anim

-0.67

achu

-0.67

Attempts

-0.66

armor

-0.64

 edits

-0.64

POSITIVE LOGITS

 nuisance

1.00

 fruitful

0.95

lot

0.92

 worthwhile

0.91

versive

0.90

 boon

0.88

 handy

0.87

 cornerstone

0.87

 huge

0.86

bit

0.85

Activations Density 0.278%

instances where something is described as being a good or beneficial thing

indefinite articles and the general concept of "something" or "one" as it relates to various contexts

No Comments

No Known Activations