INDEX

Explanations

terms related to reporting or documenting events

oai_token-act-pair · gpt-3.5-turbo

phrases related to sources and reporting in academic or news contexts

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

istries

-1.10

ierce

-0.85

visible

-0.77

incial

-0.77

atri

-0.74

arel

-0.73

idity

-0.73

hit

-0.72

rontal

-0.72

estern

-0.72

POSITIVE LOGITS

 evidenced

0.77

 unfolds

0.69

by

0.66

 guid

0.65

 narrated

0.65

 Highlands

0.64

AST

0.64

 indent

0.63

 deviations

0.63

 develops

0.62

Activations Density 0.041%

terms related to reporting or documenting events

phrases related to sources and reporting in academic or news contexts

No Comments

No Known Activations