INDEX

Explanations

phrases expressing possibility or likelihood

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the likelihood or probability of events occurring

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

 Chains

-0.74

Cause

-0.70

ummies

-0.65

imb

-0.64

 Crate

-0.64

hes

-0.63

ventures

-0.63

obe

-0.61

utch

-0.60

VB

-0.60

POSITIVE LOGITS

 overlap

0.82

 unanim

0.80

 disagreement

0.72

 impat

0.68

 denying

0.68

 appetite

0.68

 inconsistency

0.67

othal

0.67

 temptation

0.67

 indication

0.66

Activations Density 0.138%

phrases expressing possibility or likelihood

phrases indicating the likelihood or probability of events occurring

No Comments

No Known Activations