INDEX

Explanations

phrases related to physical activities or tasks

oai_token-act-pair · gpt-3.5-turbo

repeated conjunctions and connectors indicating continuity or relation in 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

 Â®

-0.68

.*

-0.67

panel

-0.65

NULL

-0.64

!:

-0.64

Card

-0.63

ential

-0.62

Fed

-0.61

("

-0.60

Adds

-0.59

POSITIVE LOGITS

romeda

1.07

 blah

0.95

 stuff

0.93

 THEN

0.91

 yeah

0.91

 secondly

0.87

 maybe

0.87

 everybody

0.86

 thats

0.86

 hopefully

0.85

Activations Density 0.409%

phrases related to physical activities or tasks

repeated conjunctions and connectors indicating continuity or relation in context

No Comments

No Known Activations