INDEX

Explanations

phrases indicating a sequence of events

oai_token-act-pair · gpt-3.5-turbo

references to events or statements occurring in relation to previous actions or developments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

thood

-0.73

 Chosen

-0.68

orce

-0.64

alpha

-0.62

ogenic

-0.60

 Beta

-0.60

bones

-0.60

 cared

-0.59

crew

-0.59

 Experiment

-0.58

POSITIVE LOGITS

 Canaver

0.78

 Hollande

0.72

Bloomberg

0.70

HCR

0.69

Gov

0.69

amsung

0.69

 PRESS

0.69

quickShipAvailable

0.68

terday

0.66

 backdrop

0.64

Activations Density 0.184%

phrases indicating a sequence of events

references to events or statements occurring in relation to previous actions or developments

No Comments

No Known Activations