INDEX

Explanations

long verbs related to actions or decisions

oai_token-act-pair · gpt-3.5-turbo

verbs indicating actions or processes related to contributions or changes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 grouping

-0.72

 succeeding

-0.70

 attaching

-0.68

 unveiling

-0.67

 requesting

-0.66

 wounding

-0.66

cia

-0.66

 bursting

-0.64

 confir

-0.62

jet

-0.62

POSITIVE LOGITS

lass

0.80

redients

0.80

edge

0.66

GGGGGGGG

0.62

 directions

0.60

tons

0.60

Edge

0.59

 rods

0.59

 irrelevant

0.59

allery

0.58

Activations Density 0.367%

long verbs related to actions or decisions

verbs indicating actions or processes related to contributions or changes

No Comments

No Known Activations