INDEX

Explanations

instances of words related to people

oai_token-act-pair · gpt-3.5-turbo

mentions of "people."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Accessory

-0.82

inventoryQuantity

-0.77

tnc

-0.77

NES

-0.75

Emer

-0.70

 srfAttach

-0.69

Stre

-0.69

COMPLE

-0.69

SPONSORED

-0.68

 Lank

-0.68

POSITIVE LOGITS

 smugglers

1.01

who

0.91

folk

0.86

opausal

0.74

iuses

0.73

hood

0.72

bara

0.71

gling

0.71

 misunderstood

0.70

 else

0.70

Activations Density 0.117%

instances of words related to people

mentions of "people."

No Comments

No Known Activations