INDEX

Explanations

phrases related to an entity or organization with the suffix "ent" or "w"

oai_token-act-pair · gpt-3.5-turbo

elements related to entwinement and connections, particularly focusing on names and relationships

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 honour

-0.72

 knees

-0.67

 goodwill

-0.67

 honor

-0.65

jri

-0.65

 uneven

-0.64

 peac

-0.64

 inconsist

-0.62

 favour

-0.61

 imitation

-0.61

POSITIVE LOGITS

lement

0.95

lements

0.82

ption

0.80

ciating

0.79

ordes

0.78

uries

0.77

 Lauder

0.77

lees

0.76

Ã©s

0.75

ement

0.75

Activations Density 0.073%

phrases related to an entity or organization with the suffix "ent" or "w"

elements related to entwinement and connections, particularly focusing on names and relationships

No Comments

No Known Activations