INDEX

Explanations

specific terms related to a political context, such as names and policies

oai_token-act-pair · gpt-3.5-turbo

references to political figures and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

hran

-0.76

ileaks

-0.71

ortment

-0.71

ingham

-0.71

itle

-0.71

estate

-0.70

ering

-0.69

umi

-0.69

 Sakuya

-0.69

achi

-0.68

POSITIVE LOGITS

cffffcc

0.74

forth

0.73

utive

0.70

olds

0.70

served

0.69

ãĥ«

0.69

oute

0.66

cers

0.66

aded

0.66

Charge

0.65

Activations Density 0.077%

specific terms related to a political context, such as names and policies

references to political figures and related terms

No Comments

No Known Activations