INDEX

Explanations

mentions of people being aware of or hearing about specific information or events

oai_token-act-pair · gpt-3.5-turbo

conversational phrases that engage the audience

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

UTC

-0.68

Reloaded

-0.67

 chapel

-0.65

 Ambro

-0.63

Bridge

-0.60

Els

-0.60

 Sabb

-0.59

 unspecified

-0.58

Render

-0.58

 hosp

-0.57

POSITIVE LOGITS

ptives

0.77

 wondering

0.76

swer

0.74

WER

0.74

ourt

0.73

hin

0.70

lance

0.69

inea

0.68

 reading

0.67

 WATCHED

0.66

Activations Density 0.393%

mentions of people being aware of or hearing about specific information or events

conversational phrases that engage the audience

No Comments

No Known Activations