INDEX

Explanations

suggestions or recommendations

oai_token-act-pair · gpt-3.5-turbo

uses of the word "suggested."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

respective

-0.74

vin

-0.70

ccording

-0.68

aliation

-0.68

mb

-0.67

brance

-0.66

kai

-0.65

through

-0.65

fare

-0.65

WHERE

-0.64

POSITIVE LOGITS

 suggestions

0.80

ible

0.76

ibility

0.69

IONS

0.68

suggest

0.68

 hints

0.68

ively

0.67

heny

0.67

antle

0.65

ion

0.65

Activations Density 0.036%

suggestions or recommendations

uses of the word "suggested."

No Comments

No Known Activations