INDEX

Explanations

terms related to wide-reaching effects, actions, or conditions

oai_token-act-pair · gpt-3.5-turbo

terms related to broad associations or generalizations within a specific context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

oleon

-0.83

wang

-0.79

pload

-0.79

mass

-0.77

irlf

-0.76

mob

-0.76

GUI

-0.75

mat

-0.74

arte

-0.74

duct

-0.74

POSITIVE LOGITS

 condemnation

0.77

 closure

0.70

 inspection

0.69

 interventions

0.69

 emergency

0.67

 enhancements

0.67

 closures

0.67

 solutions

0.66

 continuum

0.65

omn

0.65

Activations Density 0.137%

terms related to wide-reaching effects, actions, or conditions

terms related to broad associations or generalizations within a specific context

No Comments

No Known Activations