INDEX

Explanations

actions followed by subsequent actions or consequences

oai_token-act-pair · gpt-3.5-turbo

instances of the word "then" used in a sequential or procedural context

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

MpServer

-0.67

borgh

-0.64

bone

-0.64

toe

-0.63

sth

-0.63

str

-0.62

orde

-0.62

oret

-0.61

bands

-0.61

faced

-0.61

POSITIVE LOGITS

 proceeded

1.04

 conclud

0.87

 secondly

0.86

 proceed

0.83

 succumb

0.81

tyr

0.78

 promptly

0.77

 ingred

0.75

 exha

0.75

 proceeds

0.74

Activations Density 0.050%

actions followed by subsequent actions or consequences

instances of the word "then" used in a sequential or procedural context

No Comments

No Known Activations