INDEX

Explanations

phrases related to positions or roles in relation to events or situations

oai_token-act-pair · gpt-3.5-turbo

terms associated with performance and public events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Helpful

-0.88

,,,,

-0.82

zzle

-0.72

 ..........

-0.72

ynthesis

-0.71

theless

-0.71

 Effect

-0.69

 Subtle

-0.67

 Quote

-0.66

_>

-0.65

POSITIVE LOGITS

querque

0.77

 suprem

0.75

 hunter

0.75

 regist

0.75

 retailer

0.75

 worker

0.74

 powerhouse

0.74

 workforce

0.73

 insurer

0.73

 academy

0.73

Activations Density 0.787%

phrases related to positions or roles in relation to events or situations

terms associated with performance and public events

No Comments

No Known Activations