INDEX

Explanations

phrases indicating a shift to a new topic or action

oai_token-act-pair · gpt-3.5-turbo

instances of future intentions or commitments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

quickShipAvailable

-0.65

MpServer

-0.65

Gad

-0.63

 soaring

-0.63

Creat

-0.62

 constituted

-0.60

pires

-0.59

 inflamm

-0.58

 objectionable

-0.58

 Sequ

-0.58

POSITIVE LOGITS

 gladly

1.20

 assume

0.99

 presume

0.97

try

0.96

 endeavour

0.96

 continue

0.94

 reiterate

0.94

 discuss

0.94

 ende

0.91

 revisit

0.90

Activations Density 0.113%

phrases indicating a shift to a new topic or action

instances of future intentions or commitments

No Comments

No Known Activations