INDEX

Explanations

years in the format 'YYYY'

oai_token-act-pair · gpt-3.5-turbo

numerical values and references to years in the context of rankings or lists

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

reci

-0.66

advertising

-0.63

aten

-0.61

stakes

-0.61

idding

-0.61

TABLE

-0.61

GREEN

-0.59

RG

-0.59

packing

-0.59

oros

-0.58

POSITIVE LOGITS

nesia

0.65

uther

0.64

 Lawn

0.63

pee

0.62

gae

0.60

aja

0.60

yrinth

0.59

 lunar

0.59

Â½

0.59

 Andromeda

0.58

Activations Density 0.210%

years in the format 'YYYY'

numerical values and references to years in the context of rankings or lists

No Comments

No Known Activations