INDEX

Explanations

language related to setting up and configuring software or hardware

oai_token-act-pair · gpt-3.5-turbo

phrases related to the action of initiating or increasing processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 scapego

-0.75

Ø¯

-0.68

ihadi

-0.68

 Geneva

-0.66

furt

-0.66

 sworn

-0.65

 rejo

-0.64

çĶ°

-0.64

Synopsis

-0.63

 Emanuel

-0.63

POSITIVE LOGITS

river

1.02

olicy

0.96

rights

0.87

grades

0.87

grading

0.82

np

0.80

stairs

0.79

graded

0.76

ministic

0.73

dates

0.73

Activations Density 0.068%

language related to setting up and configuring software or hardware

phrases related to the action of initiating or increasing processes

No Comments

No Known Activations