INDEX

Explanations

instances of the word "follow up" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to follow-up actions or communications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 nerv

-0.65

 Scotia

-0.65

 ILCS

-0.62

 Turks

-0.61

 DRAG

-0.61

 Dictionary

-0.60

 coerced

-0.59

ãĥĨãĤ£

-0.58

 Known

-0.58

 Fant

-0.58

POSITIVE LOGITS

up

1.11

ings

1.08

ership

0.99

ups

0.99

pend

0.93

overs

0.90

through

0.89

downs

0.86

upp

0.83

word

0.83

Activations Density 0.020%

instances of the word "follow up" in various contexts

references to follow-up actions or communications

No Comments

No Known Activations