INDEX

Explanations

instructions or steps in a recipe

oai_token-act-pair · gpt-3.5-turbo

phrases indicating continuation or duration in instructions

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

VP

-0.77

egu

-0.77

pour

-0.73

foreseen

-0.71

spons

-0.70

cript

-0.70

hazard

-0.69

yre

-0.69

 sponsored

-0.68

-0.67

POSITIVE LOGITS

 Bones

0.73

 eternity

0.68

resso

0.68

adle

0.65

ithub

0.64

izont

0.63

 Carth

0.62

atcher

0.61

renches

0.61

Lob

0.60

Activations Density 0.025%

instructions or steps in a recipe

phrases indicating continuation or duration in instructions

No Comments

No Known Activations