INDEX

Explanations

commands or instructions to perform specific actions

oai_token-act-pair · gpt-3.5-turbo

modal verbs indicating future actions or events

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

quished

-0.75

Newsletter

-0.73

ocular

-0.70

JD

-0.69

foreseen

-0.67

gn

-0.66

Rex

-0.64

 Romeo

-0.64

Plan

-0.63

licted

-0.63

POSITIVE LOGITS

be

1.13

 gladly

1.07

 suffice

1.06

 doubtless

1.04

 probably

1.04

 undoubtedly

1.04

 definitely

1.01

 ALWAYS

0.97

 surely

0.97

 likely

0.96

Activations Density 0.185%

commands or instructions to perform specific actions

modal verbs indicating future actions or events

No Comments

No Known Activations