INDEX

Explanations

events or milestones that are part of a series or sequential occurrences

oai_token-act-pair · gpt-3.5-turbo

occurrences of initial events or milestones

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

contin

-0.79

sports

-0.73

lich

-0.70

etheless

-0.67

vor

-0.66

pmwiki

-0.65

tools

-0.64

ĺħ

-0.63

prov

-0.63

rench

-0.62

POSITIVE LOGITS

 EVER

0.80

 introdu

0.80

 ever

0.77

 imaginable

0.74

 arrives

0.71

 ascent

0.66

 arrive

0.65

iaz

0.65

 landfall

0.63

 Alto

0.63

Activations Density 0.352%

events or milestones that are part of a series or sequential occurrences

occurrences of initial events or milestones

No Comments

No Known Activations