INDEX

Explanations

references to documents or information that has been acquired or retrieved

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the acquisition of documents or information

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

heartedly

-0.68

odd

-0.68

gaard

-0.67

hov

-0.67

cas

-0.67

este

-0.62

mans

-0.61

psc

-0.60

brace

-0.59

border

-0.58

POSITIVE LOGITS

CLASSIFIED

0.88

racuse

0.87

andom

0.85

Materials

0.83

 unfocusedRange

0.81

cffff

0.81

OGR

0.75

>]

0.75

 permission

0.74

 obtained

0.73

Activations Density 0.022%

references to documents or information that has been acquired or retrieved

phrases indicating the acquisition of documents or information

No Comments

No Known Activations